ARK • Disrupt

In the last ten days, BlackRock, the world's largest asset manager, filed¹ an application for a Bitcoin ETF with the Securities and Exchange Commission (SEC) to be listed on the Nasdaq. One week after the SEC sued² two of the industry’s largest exchanges—Coinbase and Binance—for offering unregistered securities, the investment behemoth selected Coinbase as its ETF custodian.

While it looks much like previous ETF filings, BlackRock worked with Nasdaq to distinguish its application with a unique surveillance-sharing agreement³ designed to prevent the risk of bitcoin-related market manipulation. Based on our research, however, other applicants will be able to amend their filings with similar agreements at little cost.

In our view, BlackRock’s decision to file for a Bitcoin ETF signals that large institutional players are positive on the long-term outlook for the digital asset. Now that BlackRock has engaged with the SEC, other issuers like Invesco, WisdomTree, Valkyrie, and Bitwise⁴ are throwing their hats back into the ring.

On April 25, 2023, ARK and 21 Shares filed an application with the SEC for a Bitcoin ETF that now is the only one ahead of BlackRock’s. In our view, BlackRock’s entry is a promising step toward the acceptance of a Bitcoin ETF and could be a significant turning point in bitcoin's path to institutional acceptance.

Last week, Tesla released its compute capacity forecast,⁵ making clear that it will be not only one of the top electric vehicle (EV) manufacturers but also one of the most important AI companies in the world. Tesla is projecting that by October 2024 its compute capacity will hit a staggering 100 exaflops, ~20x more than the 4.5 exaflops it disclosed during AI Day⁶ last September. Operating at nearly 100% utilization on its current AI training infrastructure, Tesla is scaling its compute capacity aggressively to capitalize on its significant data advantage and win the race to autonomous driving.⁷

In other words, Tesla may become one of the largest AI training companies in the world. In the two years ending October 2024, Tesla intends to scale its capacity from ~14,000 Nvidia A100 chips to the compute equivalent of ~300,000 A100 chips, as shown below. For perspective, with ~80% of the market in the past year,⁸ Nvidia shipped⁹ an estimated ~350,000 SXM server GPU units.

Source: Tesla AI. 2023. Data as of 2023. Tesla Dojo is a supercomputer designed by Tesla for computer vision video processing and recognition to train its machine learning models to assist its Full Self-Driving system. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Projections may not be realized.

While it probably will continue to purchase hardware from Nvidia, Tesla is likely to shift increasingly to its own hardware—Dojo—to reach 100 exaflops of capacity at ~3X the performance of Nvidia’s A100 in October 2024. That said, because Nvidia’s latest chip—the H100—already is 4-6x more performant than the A100, Tesla will have to iterate and advance its hardware rapidly if it is to dominate the race to full autonomy in the transportation space.

Last month, Cruise announced¹⁰ that it had gathered 2 million miles of driverless data, the second million at a rate ~5x faster than the first million hit only three months earlier.¹¹

Compared to the ~2.4 years¹² Waymo needed to gather 1 million “Rider-Only” miles of data after launching¹³ its service, Cruise’s achievement is impressive.

That said, according to our research, Tesla’s Full Self-Driving (FSD) platform is gathering 1 million miles of real-world driving miles every ~14 hours as of April 2023, as shown below. Thanks to its ability to provide over-the-air updates to the ~4 million vehicles in its fleet, ~400,000 of which currently have FSD, Tesla has a staggering competitive advantage on the road to full autonomy.

[1] Kharif, O. and Greifeld, K. 2023. “Bitcoin Jumps on Speculation BlackRock ‘May Know Something.’” Bloomberg.
[2] Stempel, J. et al. 2023. “US tightens crackdown on crypto with lawsuits against Coinbase, Binance.” Reuters. 
[3] Allison, I. and Braun, H. “BlackRock May Have Found Way To Get SEC Approval For Spot Bitcoin ETF.” Coindesk.
[4] Light, J. 2023. “Hope for a Bitcoin ETF Gain Steam. 2 Reasons It Might Get Approved.” Barron’s. 
[5] Tesla AI. 2023. “And will be trained on enormous amounts of compute.” Twitter. 
[6] Tesla. 2022. “AI Day.” 
[7] Maguire, D. 2023. “Tesla's data capability is dwarfing peers.” Twitter. 
[8] Mujtaba, H. 2023. ““GPU Shipments Continued To Decline In Q1 2023: NVIDIA at 84%, AMD at 12%, Intel at 4% Market Share.” WCCF TECH.
[9] Morgan, T.P. 2023. “Just How Big Are Nvidia’s Server And Networking Businesses?” The Next Platform.
[10] Cruise. 2023a. “2,000,000 driverless miles! As sweet as the first million, just 5X faster 🙌.” Twitter. 
[11] Cruise. 2023b. “One Million Driverless Miles.” Cruise.
[12] Waymo. 2023. “First Million Rider-Only Miles: How the Waymo Driver is Improving Road Safety.” Waymo.
[13] Waymo. 2020. “Waymo is opening its fully driverless service to the general public in Phoenix.” Waymo. 
[14] Tesla. 2023. “2023 Annual Shareholder Meeting”. YouTube. 
[15] Cruise. 2023a. “2,000,000 driverless miles! As sweet as the first million, just 5X faster 🙌.” Twitter. 
[16] Waymo. 2023. “First Million Rider-Only Miles: How the Waymo Driver is Improving Road Safety.” Waymo.